Research On Message Transmission Of Opportunistic Networks Base On Community

In traditional ad hoc networks, the mobility of nodes or obstruction of signal always leads to the failure of message transmission. To deal with such situation, researchers proposed opportunistic networks. In opportunistic networks, the nodes take advantage of nodes’mobility to transmit messages, and change the conventional transmission pattern of "store-forward" to "store-carry-forward". In this thesis, we focus on the current substantial mobile devices which form a network of Pocket Switched Networks (PSNs), combine the nodes’behavior of social aggregation, and study the routing mechanism. We carry out intensive analysis and research, and finally get some meaningful achievements.This thesis will introduce the concept, background, real-life scenarios and key technologies in opportunistic networks, and we make a deep research on several typical routing protocols. According to some routing protocols, we can see that researches on opportunistic networks are mainly focus on improving the network performance, including improving delivery rate, reducing transmission delay and network overhead. Although some schemes can perform well in some specific scenarios, they do not consider the impact of human relationships on routing. If applied to PSNs, the schemes will lead to long delay and low delivery ratio. PSNs are human-centric because they follow the way humans come into contact. Human behavior always shows a character of aggregation which called "community". To sum up the previous community-based routings, they have the following problems.1. Community is static which can not reflect the real network variation.2. Message transmission is global, but the fact is that a node could decide its next hop based on local information.3. They did not consider the function of community forwarding.To address the problems above, we take full consideration about the social relationship of aggregation, the heterogeneous character within and between groups, and propose a Community-driven Hierarchical Message Transmission Scheme (CHMTS). The hierarchical thought refers to different message transmission in intra-community and inter-community. In the scheme, we determine the connections between nodes by the delivery predictability which generated by Prophet. We use the Extremal Optimization algorithm and set a metric of clock tick to dynamically divide the network into communities. In intra-community transmission, we transmit messages based on Prophet Algorithm which can find a better node for transmission, while in inter-community transmission we transmit messages based on communication weight.In order to evaluate CHMTS, we chose the ONE as simulation tool to make simulation and performance analysis. We verify the advantages and disadvantages of our scheme under Working Day Movement Model, and take comparison with the other classical protocol of Epidemic and Prophet. Experiments on the ONE show that CHMTS can raise the delivery ratio, improve the efficiency of forwarding messages, and reduce the delivery delay to some extent.